Official websites use .gov
A .gov website belongs to an official government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.
Event Date/Time: June 9, 2026 | 1:00 P.M. – 3:30 P.M. EDT
The NIST NCCoE is hosting a webinar to showcase recent work demonstrating the use of the NIST Privacy-Enhancing Technologies (PETs) Testbed and the NIST Dioptra platform to enhance AI cybersecurity and privacy.
During this webinar, the project team will share their work to secure AI model training, describe the current use cases focused on genomic data, and request feedback to help define additional use cases and data types for future demonstrations.
Background:
Analyzing data (including training AI models) using sensitive data sources without adequate privacy protections can lead to the inadvertent disclosure of private information.
Privacy-Enhancing Technologies (PETs), such as differential privacy and federated learning, can enable secure data analysis and model training while protecting private information. The PETs Testbed is designed to evaluate these technologies. In 2025, the project conducted a public red teaming exercise to assess risks from real-world privacy attacks while analyzing genomic data. The team is seeking input to help define additional use cases and data types for scalable privacy solutions.
Dioptra is a software test platform for facilitating various types of evaluations, including attacks and defense for machine learning systems. The NCCoE is developing new features for Dioptra to improve usability and aid in the design of experiments including examining how it can be used to train machine learning models with genomics data.
Event Details:
This webinar will provide a detailed overview of the NCCoE’s work on the PETs testbed and development of the Dioptra software testing platform, including:
- Details on the NCCoE’s Privacy-Enhancing Technologies (PETs) Testbed, including results from the Genomics Privacy-Preserving Federated Learning (PPFL) Platform Red Teaming Event.
- Discussion of the genomics team’s collaboration with the NIST Dioptra software testing platform for AI model training and provenance tracking.
Participants will hear directly from the project team and collaborators and will have opportunities to ask questions and provide feedback to the NCCoE.
Register Now
Register for this webinar to learn more and participate in audience Q&A. Attendance for this event is limited to 500 virtual participants.